Make PMID extraction stricter

[dstillman]

https://forums.zotero.org/discussion/69134/will-add-by-identifier-gobble-up-anything-and-output-random-new-items

[dstillman]

Still need to fix this. We shouldn't be matching on numbers bounded by hyphens.

https://forums.zotero.org/discussion/71740/usurp-isbn

[mrtcode]

The current PMID matcher is very permissive - extracts numbers separated by non-digit characters i.e. from a1%2g3 it extracts 1,2,3. I think we should only allow commas and spaces between PMID identifiers.

[dstillman]

Yeah, I agree with that. (Should be whitespace generally, so that you can paste multiple PMIDs separated by newlines.)